A Small Change that Could Make Ycombinator (and you?) More Successful

Ycombinator has become a historic institution. The amount of value and positive change they have created in the world as a private company is almost unmatched, competing mainly with nation-states and some of the oldest institutions in western society... which is a pretty crazy accomplishment in such a short period of time (currently +$350 Billion and counting).

Who/what is Ycombinator? Ycombinator or YC is an early investment (Seed and Series A) fund started by Paul Graham. They differentiate from normal seed investors in that they are run entirely by former successful startup founders and that they provide resources/advice to new startups

Arguably, Ycombinator's main contribution to the world is a growing corpus of philosophy and perspectives (now as artifacts in the form of essays, video essays, and tutorials) that teach entrepreneurial skills/ideas to a world that has become increasingly hostile to new ventures. They are teaching a new generation skills that are valuable both to the individual and to the world as a whole.

However, there is one thing that they are disseminating that could be improved, and I think it's worth raising the issue:

YCombinator's (grossly oversimplified) recommendation for how to start a startup:

  1. Come up with an idea for a product
  2. Build a product MVP based on a defined problem
  3. Talk to users - Start selling/marketing/distributing your solution
  4. Establish product-market fit
  5. Iterate and grow.

While this advice isn't bad per se, it is optimizing for the Venture Capital lifecycle, not for the success of a company.

Big, Quick, and Profitable

All tech venture capital, whether it's a Softbank or an angel investor, assumes a high failure rate and requires successes to be BIG, QUICK, and PROFITABLE. (You can read an excellent breakdown of how VC functions here: Article). Paul Graham himself even explains that Ycombinator was in his mind "such a great hack" in his essay on why they started YC, he continues by saying:

"There are thousands of smart people who could start companies and don't, and with a relatively small amount of force applied at just the right place, we can spring on the world a stream of new startups that might otherwise not have existed."

Putting this another way: The Ycombinator model has access to thousands of top ideas, talent, and startups to choose from because they created a system for developing those startups.

However, the system that Ycombinator created also REQUIRES large numbers of product-focused businesses (input) built by talented individuals for their system to produce a limited number of large successes (output). And therein lays the crux for a founder that wants to have a successful business. High growth, straight-to-product companies, have a high rate of failure. And if you're trying to optimize your business to be a "Ycombinator type" company you might be optimizing your business towards a greater likelihood of failure.

From Ycombinator's perspective, the failure/success of an individual business is irrelevant, they have economies of scale on their side. They're backing 200 high-quality startups every summer and winter and have been doing so for over a decade. The only thing they need is that their successes come quickly and are large enough to eclipse any failures that may happen.

But that's how the game works right?

Well,actually there's another way...

A (grossly oversimplified) recommendation for how to start a startup that is less likely to fail:

(for lack of a better term I'll call it a "Progressive Model")

  • Custom Solution First - Find a client that has a problem and will pay you (hopefully well) to solve it. Make sure to select the right problem (and right client) so that you can work on this problem after the initial client.
  • Solution-as-a-Service - Take that solution and apply it to a similar problem for a different client. Get paid while figuring out the commonalities of the solution. You now have a model you can apply to multiple clients. Continue doing this: getting paid to refine what will be your future product.
  • Productized Solution - Eventually, generalize your solution in the form of a product then increase the audience size.

Said more concisely:

The "Ycombinator Model":
Idea => Build a thing =>Success?

The "Progressive Model":
Find existing problem => Build a Custom Solution => Build a Service from the Solution => Build a Product from the Service => Success.

A Subtle Nuance

If you've read any of the material that Ycombinator puts out this isn't so very different from their recommended process. But there are a few subtle nuances:

First, the Ycombinator model wants you to go straight to having a product. From their perspective, it shortens the timeline for having a successful outcome but puts a stronger emphasis on starting with a great "idea", "insight", or arbitrage opportunity.

Conversely, if you start with a custom solution you are afforded an opportunity to explore the problem space, and you'll get paid to do this. This also changes the focus from being "what is our problem" to "is this the right problem"... which is very powerful. When you start with a problem that someone is willing to pay for you, there is no doubt that the solution is valuable, which may or may not be true when you go straight to product - the lingering question then becomes "can you generalize it?" and make it valuable for others.

There are many potential downsides of course to the "Progressive Model." Moving between phases (custom > service > product) is very sticky, there are lots of companies out there that can't manage to go from one to the next. Another big one, which could explain Ycombinator's strategy, is that you will almost certainly lose in any markets where there are strong network effects or a first-mover advantage. Of course, typically the prevalence of these factors are vastly exaggurated.

Lastly, there is value to failing quickly. While no one likes to fail, if you're going to fail, arguably, its best to do it quickly and get back to life without the of being a founder.

"Do as I Say" or "Do as I Do"?

Ycombinator is obviously doing a great job at putting out high-quality tech startups into the world, but is it possible that they are limiting themselves? At the scale that they currently operate, does it make sense to cling to their "hacker ethos" and outright ignore high quality service companies, consulting companies, or sales companies?

...and on the other side of the equation, what about the founders of these companies? How many founders had/ have all the pieces for a great business, but didn't make it because they were locked into building a product before they had the other pieces of the business figured out? How many potentially amazing founders are sitting around in corporate gigs because they don't have the "billion-dollar product idea." How much value could be created in the world if we had more people hunting for opportunities?

I think the underlying question to all the above is:

"What is the best way to build a business when you are starting from scratch?"

...and should YOU use the "Ycombinator Model" or a "Progressive Model?"

Obviously, my primary claim is that both models are valid, but like pretty much all useful advice in life, it really depends on your situation. Every market will be different, every project will be different, and every team of people will be different too.

While Ycombinator has had great success with helping product-focused companies grow their business, they themselves are actually using a "progressive model" - effectively they started as a service/consulting firm (in-person advice and money in exchange for a decent chunk of equity) and have been slowly transitioning to a product over the last 10 years. Productized enhancements such as software to help founders, a content library to replace direct correspondence with YC partners, developing "work at a startup," and more features that are replacing manual services/ processes as they've been growing.

With that said, Ycombinator isn't you or your company. Nor are you Stripe, Airbnb, Cruise, or Twitch. You need to find your own way. Whether a "Progressive Model" would work for you and your company I have no idea. My only recommendation is don't leap into the fray of without considering what makes sense for your situation, and if you are an entrepreneur waiting on that "dare to be great" moment, don't wait - go find a problem that you can solve.

If you liked the article you should check out my monthly newsletter "Cocktail Napkin Math":